The present invention relates to the field of powering a system comprising multiple components, and in particular to a system and entities for DC power pooling.
Systems comprising multiple modules, such as communication systems, commonly comprise modules having on-board power supplies that are connected to a common mains. In order to provide uninterrupted operation of the system, including during failure of power mains, systems often comprise a back-up power supply, such as an uninterruptible power supply (UPS), which during a power interrupt functions to supply AC power to each of modules in the system.
The combination of modules each comprising an on-board power supply often leads to a less than optimum situation, in which the on-board power supply of some modules are operating at low utilization, while the on-board power supply of other modules are approaching or have reached maximum utilization. Power supply longevity is at least partially a function of the utilization rate, in which typically a highly utilized power supply begins to increase in temperature. This increase in temperature leads to a shortened life for the power supply.
Certain modules, for example an Ethernet switch having power over Ethernet functionality, may have a need for additional power above that available from the on-board power supply. Prior art systems require a dedicated additional power supply to be added as a module, feeding the additional required power. Such a dedicated additional power supply is required despite the on-board power supply of other modules in the system being at low utilization, thus having spare power available. Furthermore, in prior art systems, a dedicated additional power supply will typically be initially underutilized, and will only experience optimum utilization as the system power needs grow. This underutilized dedicated additional power supply is thus unavailable in the event that one of the other modules in the system has reached maximum utilization of its on-board power supply.
In the event that one of the modules in the system experiences an on-board power supply failure, the prior art further does not teach an arrangement for supplying power to the module in place of the local power supply. Furthermore, the prior art does not teach an arrangement in which the utilization of local power supplies is optimized.
U.S. Pat. No. 6,125,448 issued to Schwan et al. discloses a method and apparatus of powering components on a network by using a load-share technique and by using over-voltage and current limiting circuitry. Under normal operation of the power subsystem, the load will be powered directly from the power subsystem. Unfortunately, no means of optimization of overall network power is described.
U.S. Pat. No. 5,745,670 issued to Linde discloses a fault tolerant power supply system including a plurality of nodes coupled to a common power distribution bus. Under normal operation of the power subsystem, the load will be powered directly from the power subsystem, and excess power is available to be supplied to the bus. Upon failure of the local power supply, bus power is supplied under certain conditions. No means of optimization of overall network power is described, and no means of centralized control of individual local power supplies exist.
IEEE 1394 specification, “IEEE Standard for a High Performance Serial Bus”, IEEE Std 1394-1995, Aug. 30, 1996, describes a high speed serial bus that includes the capability for sourcing power from one “node” to another over a power bus coupling the nodes. This power sourcing capability introduces potential complexities into the process of configuring the power source/sink relationships between a set of nodes or systems, such as those coupled by a 1394 specification compliant bus. For example, at any given time, one node should be providing or sourcing power and the remaining nodes should either consume power as a power sink, power themselves, or act as a power “conduit” distributing power from the power source to nodes coupled to the power distribution bus or cable (but not directly coupled to the power source). Such a layout does not teach an arrangement or a means allowing for optimization of overall network power.
U.S. Pat. No. 6,539,484 issued to Cruz describes an electronically configurable physical arrangement of power transistors. The arrangement is configurable under externally derived electrical signals to: sink power to a node from a power bus segment; source power from the node to a power bus segment; and distribute power through the node. Such an arrangement allows flexibility and power sharing, however it does not optimize overall network power.
There is therefore a need for an arrangement in which the utilization of local power supplies is optimized.